Health & Medicine

Parkinson's voice initiative uses voice-based test for potential detection

This story is a part of

Human Needs

This story is a part of

Human Needs

Screen_Shot_2013_02_04_at_5.23.00_PM_969006598.png

In Max Little's talk at TED, he discussed his goal of developing a test for Parkinson's, notoriously difficult to detect, that can be conducted via cell phone. (Photo from Ted video.)

There are 6.3 million people worldwide that have Parkinson's disease, but not all of them have been diagnosed.

Player utilities

Listen to the Story.

Since the disease has no biomarkers and a simple blood test can't predict the disease, detection is difficult. But researchers at the Parkinson's Voice Initiative are developing a way to detect and monitor the disease using voice-based tests.

Max Little, a mathematician and visiting researcher at Oxford University, says using of computers and mathematical analysis of voice patterns he and other researchers have develop a promising early test for Parkinson's that only requires a patient to sing.

By comparing songs from someone who has Parkinson's and someone who doesn't, Little says, you can hear obvious differences. Sometimes the tremors characteristic of the disease show up earlier in the voice.

In cases that are harder to detect, Little says precision mathematical algorithms are used to detect onset of the disease. But, Parkinson's Voice Initiative isn't yet at the point to show whether they can detect pre-clinical symptoms via voice recordings.

"One of the things that we're doing currently, is we're looking at people who have a particular genetic pre-disposition to Parkinson's and seeing if, for those who have that gene, but don't have Parkinson's, whether or not they display the same kind of vocal symptoms that we're seeing in very early stage Parkinson's patients," he said.

The Parkinson's Voice Initiative collected 10,000 voice recordings for analysis from volunteers with and without Parkinson's.

"So far in the lab, we've been able to detect, from lab quality recordings, with about 99 percent overall accuracy, whether somebody is healthy or has Parkinson's disease," he said. "But we wanted to test this more laterally ... outside the lab — recordings that you might collect in the world."

Little says researchers are still trying to learn more about the symptoms of Parkinson's, as well as develop clinical applications for the technology.

"If we can get a technology out actually working like that in practices, using smart phones or any other kind of microphone-based device that could work online with a computer, then that's the technology that could eventually end up in clinical practice," he said.

Little started his research about seven years ago. Since then, he's become a TED Fellow and was able to use that platform to find volunteers. Little was amazed by how prepared and willing people were to help in his research.

By using devices like telephones, smart phones or the internet, Little says researchers are able to tap into vast amounts of data that are already available.

"It's a really exciting time because we're just on this cusp where technology is reaching this point where it's so cheap that we can actually feasibly do that. We could imagine doing this across whole populations now," he said.